A PCE is a functional entity that is specially responsible for path computation in a network, and may be an independent network device, or may be an apparatus or an application program in a network device. Based on a known network topology and a constraint condition, the PCE obtains, by means of computation according to a request from a path computation client (PCC), a path that meets the constraint condition.
Generally, communication is performed between a PCC and a PCE and between a PCE and a PCE by using PCE communication protocol (PCEP), a path computation request is submitted, and a path computation result is obtained. A PCC first obtains information such as location, a path computation scope, a neighbor PCE, and path computation capability of a PCE by using a PCE discovery mechanism, to select a PCE that meets a requirement, to submit a path computation request. According to the standard, generally, a Transmission Control Protocol (TCP) connection is first established between a PCC and a PCE, and then a PCEP session is established based on the TCP connection, that is, PCEP over TCP. If protection needs to be provided for the TCP connection, a TCP message-digest algorithm 5 (MD5 option may be used. However, with the development of a network technology, the TCP MD5 is replaced by a TCP authentication option (AO), and the Transport Layer Security (TLS) Encryption Protocol that provides secrecy and data integrity also emerges.
If there are various transmission protocols, how to effectively establish a PCEP session becomes an urgent problem to be resolved.